Análise e previsão de preço do leite com aplicação de modelo de regressão para tendência e comportamento sazonal

Abstract

The primary dairy supply chain in Brazil faces several challenges, especially regarding production and commercialization. Among the main obstacles are market instability and the lack of systematized information on price dynamics over time. In light of this scenario, the present study aimed to develop a predictive model using simple linear regression and, subsequently, to enhance it through multiple regression, with the purpose of comparing the performance of both in forecasting prices. For this, data from the Center for Advanced Studies in Applied Economics (CEPEA) were used, covering the period from December 2004 to April 2025. The results indicated that the multiple regression model, by incorporating seasonal variables, showed better predictive performance, with an accuracy of approximately 68%.